TPS74401: How to prevent oscillation of TPS74401

Could you give me an answer as soon as possible?

Hello Hisashi, 

I apologize for the late response.

For TPS74401, if the EN is not going to be used by an external signal, it is ok to tie it to Vin as you are doing. One way to avoid the issue us to increase the Css capacitor as it has been suggested, however, EN has to be tied as close as possible to the largest capacitance on the input to prevent any voltage droops on that line. 

May I ask what is your input voltage? 

In addition, to use the Soft Start function using Vin=Ven, then you must supply a voltage larger than 1.1V @ Vin.

Regarding the following question:

hisashi yoda said:

Also, could you please tell me specifically what kind of phenomenon it will be when the power supply oscillates?

This is somewhat hard to answer, as there could be many factors that make it oscillate, i.e RC filters, if using a switching power supply: stability, just to mention some. 

It would come down to the specifics of your scenario to understand what is going on. 

Best, 

Edgar Acosta

Hi.Edgar Acosta

Thank you for your answer.

The input voltage is + 3.3V.

This time, you answered that there is no problem connecting EN and Vin, but before, Mr. Brian answered that it sometimes oscillates when EN and Vin are directly connected. Does that mean that there is no problem?

If it oscillates, is the countermeasure to increase the capacity of the Css capacitor of TPS74401?

This time, in response to the opinion that the direct connection between EN and Vin is a problem, we added a 4.7kΩ resistor between EN and Vin following TPS74401 on the evaluation board of your TUSB7320.

Is this circuit a problem?

As for this, I am confused because in the inquiry of TUSB7320, it suddenly oscillates due to the direct connection between EN and Vin of TPS74401.

Please answer in a hurry.

Regards

Yoda

Hi Hisashi,

hisashi yoda said:

This time, you answered that there is no problem connecting EN and Vin, but before, Mr. Brian answered that it sometimes oscillates when EN and Vin are directly connected. Does that mean that there is no problem?

If Soft Start function is desired, then a sequencing of signals must be followed, this could be the issue. I can go and ask someone in the team to corroborate that VIN and EN can cause an issue if tied together, but I honestly do not see how in your application.

If you do want to use the Soft Start function then Enable must come after VIN and VBias. If this sequencing can not be fulfilled, then EN can be tied to IN and the voltage has to be greater than 1.1V, which in your case it is. 

 From your schematic I see that you are actually tying all three, Vin, Ven and Vbias.

The capacitance required on the IN and BIAS pins strongly depends on the input supply source impedance. To counteract any inductance in the input, the minimum recommended capacitor for VIN and VBIAS is 1 μF. If VIN and VBIAS are connected to the same supply, the recommended minimum capacitor for VBIAS is 4.7 μF.

Besides that, I do not see a problem on this circuit. Just have in mind that Css is for soft start, and this will increase the timing to turn on the device. Just take a note on the IN and BIAS capacitance. 

The device should not oscillate if everything in the system works. Have you tried supplying VEN and VIN individually in your application? 

It would be interesting to see how it behaves if you un-tie them. 

Best, 

Edgar Acosta

Hi.Edgar Acosta

Thank you for your answer.

I received a lot of comments, but let me simplify the story.

1)

Is the content you should be concerned about in the answer about 7.3.1 Enable, Shutdown in the TPS74401 data sheet?

2)

Our system does not use soft start. Vin = 3.3V input.

In that case, IN and BIAS are directly connected, and EN is pulled up with 4.7 kΩ with respect to Vin.

The reason is that the power is not turned on before IN and BIAS.

The data sheet confirms 7.5.2 Sequencing Requirements.

Our circuit diagram is attached.

C3458 and C3462 are not implemented.

Please pointed out if there is a problem.

Regards,

Yoda

Hi Yoda, 

Thank you for sharing your schematic. 

hisashi yoda said:

Is the content you should be concerned about in the answer about 7.3.1 Enable, Shutdown in the TPS74401 data sheet?

Yes, that is what I have been referring to. 

I went and read your original thread for TUSB7320, and I noticed you mentioned that the oscillations occurred

"About once when the power is turned on 30 times"

Do you mean this started to occur after 30 times of power on the system? 

I asked around the office, and reviewed the first schematic you showed me, and it looks fine. The Css, as discussed, is used for the Soft Start, 

and it will help on the PSRR and to reduce some noise coming into the LDO, however, it is not a requirement.  

Also, is it possible you can show scope shots of Vin, Vbias, and Vout for the TPS74401? 

Lastly, can you do an A-B-A swap? if not, can you remove the LDO from the board and test it by itself? 

Best, 

Edgar Acosta

Is C3462 needed for EN pin?

Regards,

Brian

Hi Brian, 

Not necessarily. I see that it is being used in the TUSB7320 design reference, but it is not a requirement. Below is a schematic of the EVM designed for TPS74401 as a reference as well.

Does this "rippling" occur only at the LDO? or does it appear before the LDO? 

Best, 

Edgar Acosta

it showed up on both input and output

Yoda:

      Can you take waveform without TUSB7320 ( with and without Css ) as well?

Regards,

Brian

'

Hi Brian, Yoda,

Is there any changing Loads in the system?

After reviewing the waveforms from the TUSB7320, it doesnt look like it is an issue of ringing of the LDO, but it looks like a Load Transient waveform, can this be verified? 

Best, 

Edgar Acosta